I probably won’t make any friends today, but now I’ve read one too many urbanist (many who’s ideas I usually respect) use unsound logic to support high speed rail. This argument often includes something like this: “…and furthermore, highways and airports don’t come close to paying for themselves, therefore high speed rail need not meet that hurdle either.”
Here’s some examples of the typical contradiction many usually-reasonable urbanists are making when arguing for high speed rail-
Ryan Avent in an article plagued with this pseudo-logic:
Government is going to build more capacity. Given that, what is likely to be the best investment, all things considered?
Available alternatives, as it turns out, are not all that attractive. Roads do not appear to pay for themselves any more than railways do. Receipts from the federal gas tax come close to covering federal highway expenditures, but gas is used on highways and non-highways alike, indicating that at the federal level, highways are subsidized.
I respect Mr Cowen very much, but I think it’s long past time we stopped listening to libertarians on the issue of whether or not to build high-speed rail. Who will ask whether road construction remotely passes any of the tests they’re so prepared to push on rail? And if we begin charging an appropriate fee on drivers to maintain existing roads and reduce congestion, what do they all think will happen to land use patterns and transportation mode share?
Some have emailed to ask me why I dislike Randal O’Toole so much. The main reason is because people like Avent will always be able to point to the government highway-lover from CATO and rashly proclaim all libertarians have forever lost credibility when it comes to transportation and land use. Of course, Avent’s narrow-mindedness on this topic deserves contempt too.
And Infrastructurist’s take seems to be favored by Avent, Yglesias, and others:
The construction of a high-speed rail line would require a large environmental sacrifice – construction crews would need to shape the land, poor concrete, lay the tracks, and build the stations. This work would release millions of tons of carbon dioxide into the atmosphere. But building a new highway such as Texas’ planned I-69 would require similar work and would almost certainly be just as ecologically damaging. On a somewhat smaller scale, the same can be said for new terminals or runways at airports.
In a rapidly growing state like Texas, though, a serious need for a transportation capacity upgrade is bound to arise over the next decades – especially between the state’s two biggest cities. The construction of this infrastructure would require carbon emissions on a large scale–but since we don’t yet have competing plans for highway or airport capacity expansions if the high-speed system is not built, the most meaningful question for us is the rail system’s environmental effects in operations rather than construction.
So, in other words, building either of the options, roads or rail both require “a large environmental sacrifice”, but all other options must be kept off the table, so let’s just sweep that under the rug. Yet, there is an other option to consider for those who really think something should be done about carbon: STOP WASTING MATERIAL AND ENERGY ON CONSTRUCTION OF INFRASTRUCTURE BOONDOGGLES THAT SUBSIDIZE TRANSPORTATION! That still goes double for roads and airports, where congestion and carbon emissions could be reduced through revenue-generating measures such as congestion tolling.
To me, the high-speed rail logic just doesn’t sound much different from what O’Toole might say (just interchange some words and continue to ignore facts):
Not only did the Interstate Highway System cost much less and move much more than our visionary rail network is likely to do, interstate highways have the virtue of being 100 percent paid for out of user fees. The rail system would require subsidies for pretty much all of the capital costs, most or all of the periodic rehabilitation costs, and at least some of the operating costs.
In the Infrastructurist article quoted above, Yonah Freemark smear’s Ed Gaeser’s back of the envelope critique of high speed rail (I admit, a little sloppy) with a hand-waving claim sounding eerily similar to the type Mr. O’Toole is so often criticized for making, “High Speed Rail Pays For Itself”.
He backs this bold claim with a calculation that shows how a hypothetical Dallas-Houston high speed corridor would cost $810M annually for construction and maintenance, while providing $840M in benefit. Surely, we will see many more people use this analysis as evidence to back claims that high speed rail is good without proper scrutiny. However, this analysis doesn’t even pass the O’Toole-level test of credibility, because it claims it pays for itself with 150M annually in carbon savings. I can understand making the case for analyzing carbon savings as a “benefit” to society, but one must compare against all other options for use of cash to reduce carbon emissions – at least against a no-build + congestion toll option. Just think of all the alternatives one might consider with a $810M annual budget for carbon reduction. At say $20/ton, that comes to 40 million tons a year.
On top of that, Freemark ignores all the other opportunity costs Randal O’Toole conveniently omits when claiming roads pay for themselves. These omissions include: opportunity cost of investment capital,
opportunity cost of right of way land used, legal costs of eminent domain and related delays, inevitable cost overruns, accounting for optimism bias, and interest on bonds. In my opinion, the largest of these is the opportunity costs of investment capital, which I would guess at over 15 percent compounding annually (vs a non-compounding 5% generously assumed by Glaeser and Freemark) for all costs during the 10 years (just a little optimistic?) of construction, and 8-10 percent once ridership is stabilized. Responding to Matthew Yglesias’ hasty endorsement of Freemark’s analysis as “A real cost-benefit analysis of HSR”, Tyler Cowen similarly noted:
I’m not sure what discount rates he is using but even if we put that problem aside this screams out: don’t do it. Given irreversible investment, lock-in effects, and required hurdle rates of return, this still falls into the "no" category. And that’s an estimate from an advocate writing a polemic on behalf of the idea. I’m not even considering the likelihood of inflation on the cost side or the public choice problems with getting a good rather than a bad version of the project. How well has the Northeast corridor been run?
The urbanist in me would love a vast high speed rail network – it would centralize density at rail nodes and aid agglomeration. But it just won’t be viable until government first stops wasting money subsidizing automobile and air travel. In the meantime, HSR advocates commit an intellectual fraud similar to ones Randal O’Toole and his ilk make regarding roads when they make claims that HSR can pay for itself.
If Ryan Avent is expecting to keep any credibility on infrastructure spending using these words:
In this country, we do not build transportation infrastructure for profit. Perhaps this is upsetting to the libertarians among us, but that’s how it is and how it should be.
Then, perhaps he should think twice next time he thinks of laying into Randal O’Toole for attempting to reconcile infrastructure spending using similarly shoddy arguments. Otherwise, similar to O’Toole, all the HSR advocates are saying is, “Never mind billions of dollars that must be appropriated from people of future generations. Never mind that most of those footing the bill will never ride high speed rail if they’re not fortunate enough to afford a ticket or don’t live in one of the chosen cities. Never mind the drastic effects of the construction on the environment. High speed rail would be a pretty neat thing for some cities, so ‘build baby build’.”
MarketUrbanism saysAugust 31, 2009 at 2:11 pm
I agree. First things first…
Anonymous saysAugust 31, 2009 at 1:44 pm
I just read this new report by Reason Foundation that estimates the opportunity cost of congestion in several major urban areas. The regression model is sketchy, but leaving that aside they still can't refrain from advocating for greater highway investment. Every traffic engineer knows that more roads != less congestion. Why is raising tolls so out of the question?
MarketUrbanism saysAugust 31, 2009 at 2:02 pm
Thanks for tipping us off. If you have a link, please share. I'll try to find it too.
rpitingolo saysAugust 31, 2009 at 2:06 pm
I typically agree with guys like Avent and Yglesias on these questions, but I have really been cautious about HSR. In the countries where this type of infrastructure has worked, the cities that it connects typically have excellent transit systems. Most cities in the US, frankly, do not compare. So I fear that we are getting ahead of ourselves before other important pieces of the puzzle are in place.
Rationalitate saysSeptember 1, 2009 at 12:52 am
1. I agree with your opposition to HSR. It is a boondoggle wasting taxpayers' money for little benefit at best, and a serious detriment to urbanism at worst. People will associate rail with interurban rail, and intraurban networks will feel the consequences.
2. I think it's important for honest defenders of market urbanism to point out that O'Toole is correct when he says that roads pay for themselves through the gas tax. What Ryan Advent (if my reading of the data is true) is saying is false – “highways” in actuality means all roads, big and small, locally- and federally-funded. (I believe that this figure actually also includes some police patrols.) If I'm wrong about this, let me know, but I don't think I am. That having been said, O'Toole errs in that he ignores the opportunity cost of the roads as well as the density-inhibiting regulations that essentially force people to use roads. Also, there's the fact that more-traveled roads subsidized less-traveled roads, although this goes both ways – under a more localized system of payment, all other things being equal, the subsidizing roads would be enlargened.
MarketUrbanism saysSeptember 1, 2009 at 2:31 am
I'm sorry, but nothing I've seen from O'Toole convinces me that highways or roads pay for themselves. He ignores too many significant costs in his calcs.
I'm willing to give take another look in case I missed something. Maybe you'd like to take a shot at convincing me I'm wrong with a blog post clearly defending O'Toole's numbers – at least from an accounting standpoint. And if you can successfully defend O'Toole from an investment standpoint, I'll buy you a copy of all his books. 🙂 (or maybe something with actual value)
Roadie saysSeptember 1, 2009 at 10:04 pm
“nothing I've seen from O'Toole convinces me that highways or roads pay for themselves.”
He never said roads completely pay for themselves. He said roads come much, much closer to paying for themselves than does transit. And he is absolutely correct about that.
MarketUrbanism saysSeptember 1, 2009 at 11:08 pm
Um, well, the quote I used in this post was a direct quote from Mr. O'Toole's blog:
MarketUrbanism saysSeptember 1, 2009 at 11:25 pm
here's the link to that Reason study
Wow! I just skimmed it. Unbelievable…
Daniel saysSeptember 2, 2009 at 12:33 pm
This is a fair critique, but it's not necessarily unreasonable to hold O'Toole accountable to his own principles without agreeing 100% with those principles oneself. That being said, the argument can get a little convoluted unless this difference is spelled out very clearly, and I do agree with you that coming up with numbers that attempt to show that HSR “pays for itself” is pretty sketchy.
My issue with the whole debate is that the cost/benefit analysis is way to hard to crunch for such a huge, complex program that will be implemented over the course of decades. There are simply too many factors to have a quantifiable analysis of this. Glaeser clearly acknowledges this fact, but goes ahead and runs the numbers speculatively anyway (the problem is that not everyone knows he is speculating) in his series on the NyTimes blog.
So, in my opinion, a large-scale HSR might be able to pay for itself in the long-term (20 – 30 years down the road). Since this is a longer time horizon than most private capital is willing to wait, the federal government seems like the appropriate body to carry out a vast intercity network. Alternatively, the system could be built piece-meal by private actors, but in the case of HSR it seems the system may be more efficiently built from the top-down than the bottom-up. There are just too many interlocking parts.
MarketUrbanism saysSeptember 2, 2009 at 5:51 pm
Sure. Then, they should either drop the “pays for itself” talk or lay off of O'Toole.
MarketUrbanism saysSeptember 2, 2009 at 6:38 pm
If it's too hard to quantify, then its not reasonable to expect people to pay for it. But, it's not too hard to quantify. Private companies make multi-billion dollar investments every day and they you'd better believe they do extensive due diligence to quantify the cost/benefit. They'd spend millions on research, designs, and studies and run hundreds of scenarios and sensitivities. But, what does it tell us that no private investors have already jumped in and built high-speed rail (or highways) if there are ones that “pay for itself”? It tells us that only government will get into these kind of projects without proper due diligence.
But, you say maybe the private sector doesn't have that kind of patience?
To look at this more carefully we need to understand what is different about government and private sector that would make this true?
First of all, being in the real estate development business, I know that decisions are often made for the long run. There is plenty of patient capital that looks for long holding periods of 30 yrs+. The biggest of these is pension funds, which try to match the duration of time until pensioners retirement to the duration of investments. If there was a viable high speed rail project, you can bet pension funds would want a piece of it.
The thing is according to modern finance theory, which applies to the public sector the same as the private sector, the present value of cash flow 20 years into the future is pretty small relative to the nominal amount of future cash flow. So, the longer the term, the more the present value converges. Financial always analysis takes this long-term-effect into account even if their holding period is a few years because they will want to have something valuable to sell at the end of the day. Now, if someone in government plays the trick of saying the sum of future cash flows (instead of discounted cash flow) pays for the investment, they are defying modern finance theory, and purposely (or naively) deceiving you. (this is why I encourage planning students to take at least one course in finance theory and microeconomics)
Now, let's compare that with the “patience” of government capital. Politicians don't even care about recouping the investment, because their term is only a few years, and it's likely voters 30 years from now will not care about the boondoggles of the past – like Eisenhower who is considered a hero for his highway waste. And when voters are able to rationally see the failures, the politician has probably retired already. So politicians and bureaucrats love to be the one spending on huge capital projects they can put their name on, but hate paying for maintaining past politicians' trophies. (airports are a different story because they are huge political patronage machines over the entire life-cycle of the investment) So, do you think any politician even gives a damn about whether or not a project “pays for itself” after 30 years? No, but a private company will go under if they made the decisions politicians jump on. More more importantly, private enterprise wouldn't even be able to raise the capital needed for a boondoggle, while the public sector has the ability to just tax the public more and more to pay for the failures.
MarketUrbanism saysSeptember 2, 2009 at 6:49 pm
So, in conclusion, there is no fundamental difference in the “time horizon” of public vs private capital. The difference is that only government is NEGLIGENT enough to pursue a project where the ones responsible will not have to take responsibility for its success or failure over the long run, and the burdens of failure can be shifted to the innocent.
OldUrbanism saysSeptember 3, 2009 at 12:43 am
After reading this post, I once again notice the suggestion of an alternative formulation of the “full cost” of roads (and other infrastructure). I have encountered this argument in several posts on this blog. Some of the ideas I don't think are correct, and so would like to discuss here.
The first is the issue of opportunity cost, which you give great prominence to in this post. I think it would be important to first mention in each case whether you are referring to private finance of infrastructure (e.g. roads) or social benefit-cost analysis for public provision, which is much more common. In both cases interest rates figure prominently, but are treated slightly different. The opportunity cost of investment capital is explicit in the case of private finance, it is simply factored into the interest rate. It is more difficult for social BCA. since the objective is to determine the appropriate opportunity cost. Most recent textbooks I've seen on the subject do not assume a priori that public projects simply displace private capital investment, but rather some combination of private capital and private consumption, the latter having a lower marginal rate of time preference (and hence discount rate). Since this issue is not 100% settled, the best course of action in the case of social BCA is to use sensitivity analysis for a range of values, assuming different types of discount rates.
In the case of private finance, two other factors that you mention, risk of cost overruns/uncertainty and demand uncertainty/bias can also be appropriately factored into an interest rate. In principle, the same could be done for the discount rate in social BCA, but in practice is rarely employed.
The last two factors, legal costs associated with eminent domain and opportunity costs of land, are in fact often included in typical project cost estimates for both public and private projects. The former is fairly straightforward, as it is a project-related cost. The latter, opportunity cost of land, is simply the purchase price of land (assuming it is bought at market rates through an arms-length transaction — this doesn't always happen). I have also heard some argue that the cost of existing roads should reflect the land it occupies. To me, this raises at least three questions:
1) What value would that land (presumably urban) have were it not made accessible by the provision of infrastructure. Would it have any value at all (above an agricultural use)?
2) The land typically occupied by urban infrastructure (especially roads) is not well-suited for development, and hence would have a low opportunity cost. This land often appears in long, narrow parcels, upon which little could be profitably built, especially if confronted with modern land use regulations (e.g. maximum impervious surface requirements).
3) Who is the residual claimant for the payment of the supposed opportunity cost? Local government? How would this enhance social efficiency?
Lastly, how do you arrive at an interest rate of 15 percent? This just sounds absurd. Most of the privately financed roads and other infrastructure projects I've seen recently have not faced interest rates remotely close to this level, even the risky ones. Heck, even the Channel Tunnel project did not face rates this high. As a real estate developer, have you ever been charged a rate this high for long-term debt?
There are probably a laundry list of issues one could discuss on this matter, and I've just touched on a few that caught my eye. My final point would be this: there are things we could do to better discipline infrastructure investment, but they must be tempered with realism. Even if it were possible to calculate a socially optimal cost for road users (hint: it isn't), could we actually do it? I haven't even touched on transaction costs, which become an important matter when talking about things like externality pricing or toll collection. Politically and practically, we cannot achieve social optimality. It's just not a Pareto efficient world. We should probably content ourselves with doing things that push us in the right direction and that we know will work.
MarketUrbanism saysSeptember 6, 2009 at 4:32 am
Thank you for taking the time. I welcome any input that helps me communicate more accurately and effectively. I especially appreciate input from knowledgeable, intelligent commenters like you.
I am referring to opportunity cost in a strict sense of the term: the value of a choice which is forgone in order to pursue a particular decision. Before going further, I must confess that prior to your comment, I was not familiar with “social benefit-cost analysis”. I skimmed the internet and didn’t find too many good resources on the topic. The concept still sounds a little scary to me, in a Soviet sense. What insights you can share with us?
I think you are oversimplifying quite a bit. The cost of capital for a particular investment is the weighted average cost of capital for all tranches of that investment. It includes all forms of equity, as well as debt. And even for the debt tranche, the opportunity cost of capital is not as simple as a relation to interest rate – although the concept is often incorrectly interchanged. The opportunity cost of the capital is the returns forgone to pursue a project of the same risk. Thus, I consider the Opportunity Cost of Capital a much more valid metric than interest rate for projects both public and private.
I think likelihood of default, and liquidity are the determining factors for pricing interest rate. And in infrastructure finance there must be some degree of an implied backing by government in the case of default. The value of that guarantee is huge.
Or sometimes, so-called-private infrastructure even uses government bond financing. Are these risks even a factor in government bond financing, or are bonds priced merely the risk of that government failing? And when government can always lean on its citizens to prevent default, is it really baked into the price?
But, regardless, we are not given an interest rate to work with in this example.
You are absolutely correct, and I am embarrassed for being so sloppy. Thank you for pointing that out. I will make the correction to this post accordingly.
I would like to clarify, that I stand by my posts where I rebut assertions that existing highways “pay for themselves” as we discuss below.
Of course, if there were no infrastructure, the land would have limited value. Are you suggesting I am against infrastructure or wish to destroy all of it? I am sure you are not. I am only suggesting that we take all costs into consideration, including the value of distinct pieces of land the infrastructure uses if that land were put to different use.
I don’t think this is anything close to being universally true. Let’s consider the typical urban expressway, which is the case I consider most in need of opportunity cost consideration. Most urban expressways displaced existing communities when constructed, and if the expressways were to be demolished, the land would likely return to that use or a similar alternative. At the same time, nearby property values would increase as the negative externalities of the expressway are alleviated. Thus, I consider it necessary to at least evaluate the value of the DOT land if it were to be liquidated at values similar to land in nearby neighborhoods. The actual liquidated values would likely be greater, in consideration of the alleviated externalities.
Since the land is probably owned by the DOT or some other government agency, that is the entity which I assume would receive funds from the hypothetical liquidation. To take that to it’s logical conclusion, the taxpayers who’s incomes were originally appropriated to fund the DOT would be the rightful claimants. (but we know that is impossible to expect from government)
And you’ll have to define “social efficiency”, as it’s a foreign concept to me. What is the metric of measurement of “social efficiency”?
Again, opportunity cost of capital is neither an interest rate or cost of debt. The Channel Tunnel’s original projected IRRs estimated as high as 18% in this source, so I’d peg it’s OCC in the mid teens. What was the OCC of the privately financed roads and infrastructure projects you’ve seen recently?
Let me also add that I guessed 15% for pre-stabilization and 8-10% post-stabilization, so it’d be much lower than 15% if you blended them, Maybe infrastructure isn’t analyzed in phases like that, but that’s me thinking like a developer. Your insight would be valuable.
Yes, of course I’ve been involved with projects that paid more than 15% for Mezz. debt. In fact, as high as 25%.
But, we’re talking about Opportunity Costs of Capital, not debt. So, taking into consideration equity, debt, and mezz., most projects I work on have an OCC from 12% to 20%. But, of course, none of my projects are nearly as risky as a unique, multi-billion dollar high-speed rail project subject to budget overruns, vast political interference, 10+ year time horizon, and little residual value in the case of failure.
I agree, and I wouldn’t tolerate a world where every action would be coerced for the sake of “social optimality.”
Andrew Dawson saysOctober 24, 2009 at 11:37 am
Randall O'Toole & Rationalitate are hypocrites.
Though most of those that are pushing for HSR along with better transit service are not hypocrites, since they are not against public funding for streets & roads.
rent_a_car saysJanuary 27, 2010 at 8:25 am
This is very useful indeed. The alternative would be to use sprees…
Rent a car Romania